Dynamic Structural Equation Modeling (DSEM)
Dynamic structural equation modeling (DSEM) combines multilevel modeling, time-series modeling, and SEM to estimate within-person dynamics such as autoregressive and cross-lagged effects in intensive longitudinal data.
동적 구조방정식모형(DSEM)은 다층모형, 시계열모형, 구조방정식모형을 결합하여 집중 종단자료에서 자기회귀·교차지연 효과 같은 개인 내 동태를 추정하는 방법이다.
How do I analyze within-person dynamics in daily diary or experience sampling data?
일기식 조사나 경험표집 자료에서 개인 내 변화 동태는 어떻게 분석하나요?
Dynamic structural equation modeling estimates person-specific autoregressive and cross-lagged effects at the within level and lets those effects vary randomly across persons at the between level, usually with Bayesian estimation. It fits intensive longitudinal data such as daily diaries, ecological momentary assessment, or sensor data with many time points per person.
When to use Dynamic Structural Equation Modeling (DSEM)
- Many repeated measurements per person, such as daily diaries or experience sampling
- Estimating inertia (autoregression) and cross-lagged effects within persons
- Explaining individual differences in dynamics with person-level variables
When to use something else
- Only a few waves per person: use a random-intercept cross-lagged panel model
- Only one long time series for a single unit: use a vector autoregressive model
Assumptions and requirements
- Enough time points per person to estimate within-person dynamics
- Stationarity of the within-person process (or modeled trends)
- Approximately equally spaced measurements or a time-interval adjustment
- Bayesian estimation with convergence checks
What the analysis reports
- Fixed (average) autoregressive and cross-lagged effects with credible intervals
- Between-person variances of random effects
- Within- and between-level variances
- Convergence diagnostics such as PSR (R-hat)
How to report it
The average autoregressive effect of [variable] was [value], 95% credible interval [[value], [value]], with significant between-person variance in inertia ([value]).
Related methods
- Multilevel Vector Autoregression
- Random Intercept Cross-Lagged Panel Model
- Multilevel Structural Equation Modeling
Run Dynamic Structural Equation Modeling (DSEM) on Quorant
- Data you need
- a table (CSV or Excel) with item columns (e.g. questionnaire items) and a cluster ID
- Price
- 900 credits for one analysis up to 10,000 rows (complexity level 5); larger or wider data costs more. Optional methodologist review: 3600 credits.
- Checking
- Python results are cross-checked against R reference outputs in every deployment
- Before you pay
- A free pre-analysis diagnosis checks your data and gives an exact quote; nothing is deducted until you approve it.